README
======

This repository contains the data and R script necessary to replicate the
analysis presented in Figure 1 of the following article:

Montero-Melis, G., Isaksson, P., Paridon, J. van, & Ostarek, M. (accepted/2019).
"Does using a foreign language reduce mental imagery?" Cognition.


Description of files
--------------------

### exp2_stimuli_as-shared.csv

This is the list of stimuli used in Experiment 2 of Hayakawa and Keysar (2018),
as shared with us by the authors. (Shared here with the authors' permission.)

Column explanation:

- *List*: Each of two lists used in original study for counterbalancing items.
- *w1*-*w3*: The three words that made up a given item. Their position was
  randomized in the experiment, but here w1 and w2 always represent the shape
  match and w2 and w3 the category match.
- *shape_match*/*categ_match*: clarifies which words show the shape and category
  matches, respectively.

 
### exp2_stimuli_corrected.csv

Identical to "exp2_stimuli_as-shared.csv" except for line 34, which corrects an
obvious error in the original file: the category match should be 
"chopsticks--spoon" and not "ski--spoon".


### exp2_stimuli_with_cosine_sub2vec.csv

Unique word pairs and their cosine similarities.

Columns:

- *word_pair*: word pair
- *match_type*: category or shape match
- *cosine_s2v*: cosine similarity between the two words. For details, see the
  article. Note that the word vectors are numerical representations of single
  words, therefore multiword noun phrases were excluded from the analysis.


### fig1_cosine_similarities.R

R script that generates Figure 1 in the article.
